Multipoint Genome-Wide Linkage Scan for Nonword Repetition in a Multigenerational Family Further Supports Chromosome 13Q As a Locus for Verbal Trait Disorders D

Marshall University Marshall Digital Scholar

Biochemistry and Microbiology Faculty Research

12-2016 Multipoint genome-wide linkage scan for nonword repetition in a multigenerational family further supports chromosome 13q as a locus for verbal trait disorders D. T. Truong

L. D. Shriberg

S. D. Smith

K. L. Chapman

A. R. Scheer-Cohen

See next page for additional authors

Follow this and additional works at: https://mds.marshall.edu/sm_bm Part of the Behavioral Medicine Commons, Genetic Phenomena Commons, Genetic Processes Commons, Genetic Structures Commons, and the Speech Pathology and Audiology Commons

Recommended Citation Truong DT, Shriberg LD, Smith SD, Chapman KL, Scheer-Cohen AR, DeMille MMC, et al. Multipoint genome-wide linkage scan for nonword repetition in a multigenerational family further supports chromosome 13q as a locus for verbal trait disorders. Hum Genet. 2016;135(12):1329-41.

This Article is brought to you for free and open access by the Faculty Research at Marshall Digital Scholar. It has been accepted for inclusion in Biochemistry and Microbiology by an authorized administrator of Marshall Digital Scholar. For more information, please contact [email protected], [email protected]. Authors D. T. Truong, L. D. Shriberg, S. D. Smith, K. L. Chapman, A. R. Scheer-Cohen, M. M.C. DeMille, A. K. Adams, Alejandro Q. Nato Jr., E. M. Wijsman, J. D. Eicher, and J. R. Gruen

This article is available at Marshall Digital Scholar: https://mds.marshall.edu/sm_bm/228 Hum Genet (2016) 135:1329–1341 DOI 10.1007/s00439-016-1717-z

ORIGINAL INVESTIGATION

Multipoint genome‑wide linkage scan for nonword repetition in a multigenerational family further supports chromosome 13q as a locus for verbal trait disorders

D. T. Truong1 · L. D. Shriberg2 · S. D. Smith3 · K. L. Chapman4 · A. R. Scheer‑Cohen5 · M. M. C. DeMille1 · A. K. Adams6 · A. Q. Nato7 · E. M. Wijsman7,8 · J. D. Eicher6 · J. R. Gruen1,6,9

Received: 11 May 2016 / Accepted: 22 July 2016 / Published online: 17 August 2016 © The Author(s) 2016. This article is published with open access at Springerlink.com

Abstract Verbal trait disorders encompass a wide range region in verbal trait disorders. Future studies will further of conditions and are marked by deficits in five domains refine the specific causal genetic factors in this locus on that impair a person’s ability to communicate: speech, lan- chromosome 13q that contribute to language traits. guage, reading, spelling, and writing. Nonword repetition is a robust endophenotype for verbal trait disorders that is sensitive to cognitive processes critical to verbal develop- Introduction ment, including auditory processing, phonological working memory, and motor planning and programming. In the pre- Verbal trait disorders are comorbid, developmentally asso- sent study, we present a six-generation extended pedigree ciated disorders and deficits in communication. These with a history of verbal trait disorders. Using genome-wide include clinical and subclinical disorders of speech, lan- multipoint variance component linkage analysis of non- guage, reading, spelling, and writing (Shriberg et al. 2012). word repetition, we identified a region spanning chromo- Speech sound disorders (i.e., excluding dysfluency) are the some 13q14–q21 with LOD 4.45 between 52 and 55 cM, most prevalent verbal trait disorders at preschool age, with = spanning approximately 5.5 Mb on chromosome 13. This an estimated population prevalence of 16 % at age 4 years region overlaps with SLI3, a locus implicated in reading (Campbell et al. 2003), decreasing to 3.8 % at age 6 years disability in families with a history of specific language (Shriberg et al. 1999) and 3.6 % at age 8 years (Wren et al. impairment. Our study of a large multigenerational fam- 2012). Deficits in speech frequently co-occur with impair- ily with verbal trait disorders further implicates the SLI3 ments in multiple domains. For example, 11–15 % of children with speech sound disorders at age 6 also have language disorder (a neurodevelopmental disorder that can Electronic supplementary material The online version of this article (doi:10.1007/s00439-016-1717-z) contains supplementary affect both spoken or written language; Shriberg et al. material, which is available to authorized users.

* J. R. Gruen 6 Department of Genetics, Yale School of Medicine, New [email protected] Haven, CT 06510, USA 7 Division of Medical Genetics, Department of Medicine, 1 Department of Pediatrics, Yale School of Medicine, New University of Washington, Seattle, WA 98195, USA Haven, CT 06510, USA 8 Department of Biostatistics and Department of Genome 2 Waisman Center, University of Wisconsin-Madison, Sciences, University of Washington, Seattle, WA 98195, Madison, WI 53705, USA USA 3 Department of Pediatrics, University of Nebraska at Omaha, 9 Investigative Medicine Program, Yale School of Medicine, Omaha, NE 68182, USA New Haven, CT 06510, USA 4 Department of Communication Sciences and Disorders, University of Utah, Salt Lake City, UT 84112, USA 5 Department of Speech‑Language Pathology, California State University, San Marcos, CA 92096, USA

1 3 1330 Hum Genet (2016) 135:1329–1341

1999). Additionally, children with speech disorders are at of specific language impairment (SLI) identified a linkage higher risk for reading disability, with an estimated 18 % of peak on chromosome 16q for poor performance on NWR children with speech disorders and 75 % of children with (SLI Consortium 2002, 2004). Follow-up family-based both speech and language disorders meeting criteria for and population-based association studies on NWR identi- reading disability at school age (Lewis et al. 2000). fied CMIP and ATP2C2 as candidates responsible for the Research in the genetics of verbal trait disorders was linkage signal (Newbury et al. 2009). CNTNAP2 on chro- catalyzed by the seminal studies of the KE family, a large mosome 7q35 was also associated with NWR in families extended pedigree segregating verbal dyspraxia (also enrolled in the SLI consortium study using a candidate termed Childhood Apraxia of Speech; ASHA 2007; RCSLT gene approach after it was identified as a transcriptional 2011), suggesting autosomal dominant inheritance of a sin- binding target of FOXP2 by chromatin immunoprecipi- gle gene mutation (Hurst et al. 1990). Genome-wide link- tation (Vernes et al. 2008). In addition, CNTNAP2 was age showed a signal peak on chromosome 7q31.1 (Fisher identified by fine mapping a linkage analysis signal on et al. 1998). Further fine mapping identified a point muta- 7q35 conditioned on language delay in the Autism Genetic tion in FOXP2 that resulted in a truncated protein and loss Resource Exchange sample (Alarcón et al. 2002, 2005, of function in all affected individuals, but not observed in 2008). Taken together, NWR satisfies specific testable cri- unaffected individuals (Lai et al. 2001). FOXP2 loss of teria for the objective identification of endophenotypes, function as a causal factor for verbal dyspraxia was fur- supporting NWR as a credible endophenotype for verbal ther validated in unrelated individuals with severe speech trait disorders (Glahn et al. 2014; Lenzenweger 2013). impairments similar to those in the KE family (Lai et al. With the exception of the KE family, most families with 2001; MacDermot et al. 2005) and has been cross-validated a history of language impairment show a complex pattern in a number of case studies (e.g., Rice et al. 2012; Shriberg of inheritance with subtle differences in clinical presenta- et al. 2006). Although the KE family provided an example tion within the family. In the present study, we examined of a verbal trait disorder phenotype with a typical pattern of an extended six-generation family with a complex pattern monogenic inheritance, their story is the exception rather of inheritance for verbal trait disorders. We chose NWR than the norm. In fact, verbal trait disorders are generally in this analysis because (1) it is a robust endophenotype multifactorial and associated with multiple genetic and for verbal trait disorders (i.e., speech sound disorder, lan- environmental factors (Kang and Drayna 2011; Peterson guage disorder, and developmental dyslexia); (2) is highly and Pennington 2015). heritable; (3) has a Mendelian model of inheritance (in Due to the behavioral and cognitive heterogeneity of at least one study; Wijsman et al. 2000); and (4) is stable verbal trait disorders, the use of endophenotypes—underly- throughout an individual’s lifetime, even in those who are ing phenotypic factors that are associated with or contrib- language recovered following impairment in childhood ute to the manifestation of the disorder of interest because (Bishop et al. 1996; Shriberg et al. 2009). The latter attrib- of shared genetic factors—have been critical to the genetic ute of NWR tasks is particularly important because sub- study of verbal trait disorders. One endophenotype is non- jects within this family range in age from 3 to 95 years, word repetition (NWR), which loads onto several cogni- requiring a phenotype that can be ascertained and com- tive processes critical for language-related ability includ- pared across all age groups. The present analysis provides ing auditory processing, receptive language ability, and strong support for chromosome 13q14–q21 as a locus that motor planning and programming (Dollaghan and Camp- contributes to poor performance on NWR in this extended bell 1998). NWR tasks examine the ability to process and pedigree. temporarily store a novel series of meaningless units of phonological information in short-term memory, and then verbally repeat the stimuli. Such measures, which are sen- Methods and materials sitive to but not specific for any one disorder, may be more closely influenced by genetic variation than the verbal trait Ascertainment disorder itself. NWR task performance has a strong genetic influence, We studied 62 individuals from a six-generation 90-mem- with higher concordance among monozygotic twins com- ber family of European ancestry with a history of verbal pared to dizygotic twins, and heritability ranging from 0.64 trait disorders. The family was ascertained with the assis- to 1 (Bishop et al. 1996, 2004). Furthermore, an oligo- tance of a family member. The 62 family members assessed genic-trait segregation analysis of NWR in nuclear families included 35 females and 27 males ranging in age from 3 to ascertained for reading disability estimated approximately 95 years. There is no evidence of consanguinity based on 2.4 quantitative trait loci (Wijsman et al. 2000). A family- genealogy or unexpected high kinship coefficients within based linkage analysis on individuals with a family history the pedigree.

1 3 Hum Genet (2016) 135:1329–1341 1331

Written informed consent was approved by the Uni- transcribed as correct or incorrect. The number of correctly versity of Wisconsin-Madison Institutional Review Board repeated consonants was divided by the total number of tar- (IRB). All subjects were assessed by one of two experi- get consonants. The ratio was then converted to a standard enced examiners in the participants’ homes or hotel sites in score using the reference database. five states within the continental US. All oral instructions Studies of speech-language disorders using the NRT and audio-recorded stimuli were presented at comfortable have supported its validity and reliability (e.g., Archibald listening levels based on findings from a conventional hear- and Gathercole 2006; Moore et al. 2010), including a ref- ing screening. The assessment protocol included the fol- erence sample of 95 children with typical speech and 63 lowing measures and instruments: Kaufman Brief Intelli- children with speech delay, described in a technical report gence Test-2 (Kaufman and Kaufman 2004; nonverbal and on the NRT and SRT (Shriberg and Lohmeier 2008). Find- verbal IQ), Nonword Repetition Task (NRT; Dollaghan and ings from this reference sample include psychometric data Campbell 1998), Syllable Repetition Task (SRT; Shriberg supporting the distributional characteristics of scores for et al. 2009), Goldman-Fristoe Test of Articulation-2 (Gold- parametric statistical analyses, and analyses supporting the man and Fristoe 2000; speech), Clinical Evaluation of Lan- construct validity, concurrent validity, interjudge transcrip- guage Fundamentals-Preschool-2 (Wiig et al. 2004; lan- tion reliability, and internal reliabilities of both tasks. guage), Clinical Evaluation of Language Fundamentals-4 In the present study, point-to-point percentage of agree- Screening Test (Semel et al. 2004; language), Woodcock– ment estimates ranged from 75.6 to 88 % across nonword Johnson Tests of Achievement, 3rd edition (Woodcock task and phoneme class; other validity and reliability esti- et al. 2001; reading, spelling, and writing), and question- mates were generally in the 0.70–0.85 range. The Pearson naires for parent-reporting or self-reporting medical and r coefficient between standardized scores on the two non- special educational histories and concerns. One individual word tasks in the present data was 0.66, consistent with had a composite IQ <75, but performance on the NWR the coefficient of 0.73 reported in Shriberg and Lohmeier tasks was unimpaired. To maximize genetic informative- (2008). Thus, consistent with discussion elsewhere, there ness, this individual’s NWR scores were retained for the is only moderate collinearity between the two measures of analysis. All other individuals had an IQ between 84 and NWR (Shriberg et al. 2009). 126. Sensitivity and specificity for identifying speech and language disorders using the SRT were further supported Phenotype by a second reference sample of 550 speakers, including speakers with typical speech and typical language, speech The NRT is a NWR task that consists of 16 nonwords. delay and typical language, language impairment and typi- To reduce the articulatory burden, the NRT does not con- cal speech, and speech delay and language impairment tain the most phonetically complex consonants (the “late- (Lohmeier and Shriberg 2011). Additional construct valid- 8” consonants; Shriberg 1993; Dollaghan and Campbell ity support for the SRT was presented in Shriberg et al. 1998). Nonwords ranged in length from one to four sylla- (2009), followed by a series describing SRT procedures bles (four each) with the shortest nonwords presented first to explicate encoding, memorial, and transcoding pro- and the longest last. Each repeated consonant and vowel/ cesses underlying performance on nonword imitation tasks diphthong (totaling 20 different phonemes) was later tran- (Shriberg et al. 2012). scribed as correct or incorrect by two research speech Because there is no battery of speech, language, reading, pathologists. NRT scores were calculated by dividing the spelling, and writing tests appropriate for the lifespan ages total phonemes correctly repeated by the total phoneme tar- of the present extended family, we used standardized scores gets. Ratios were then converted to age–sex standardized from either the NRT or SRT to assign a categorical pheno- scores for downstream analyses using a reference database type. Verbal trait impaired (Verbal Trait ) was defined as + of 200 typical speakers, ages 3–80 years (Potter et al. 2012; performing greater than one standard deviation below the Scheer-Cohen et al. 2013) that included descriptive statis- mean on either the NRT or SRT. Preliminary studies indi- tics for NRT and SRT scores. cated that a cutoff below one standard deviation on either The SRT is another NWR task comprised of 18 non- the NRT or the SRT was maximally sensitive and specific words that include only four of the “early-8” consonants to subjects with only mild, subclinical difficulty in one or (/b/,/d/,/m/, and/n/) and the vowel/ɑ/(Shriberg 1993). This more of the five verbal traits based on parent- and self- NWR task was designed to accommodate individuals who reported histories of children and adults. Of the 41.9 % of have incomplete phonetic inventories and/or articulatory participants in the present study who met the nonword cri- impairments. Items range in length from two to four syl- teria for a verbal trait disorder (see Table 1), 19.2 % met lables with the shortest presented first and the longest last. criteria on the NRT only, 23.1 % met criteria on the SRT The consonant responses to each recorded syllable were only, and 57.7 % met criteria on both nonword tasks.

1 3 1332 Hum Genet (2016) 135:1329–1341

Table 1 Percentagesa of affected (Verbal Trait ) and not affected (Verbal Trait ) participants in an extended family of 62 members and tests of two proportions results for each variable. Participant+ age was divided into four lifespan− cohorts

Variable Total n Verbal trait Verbal trait Tests of two proportionsb (VT ) + (VT ) − + − n % n % Z p Confidence interval Sig.c

Participants 62 26 41.9 36 58.1 1.82 0.069 0.335, 0.012 − − Gender Female 35 11 31.4 24 68.6 3.35 0.001 0.589, 0.154 * − − − Male 27 15 55.6 12 44.4 0.82 0.411 0.154, 0.376 − Age Preschool (3–5) 3 1 33.3 2 66.7 0.87 0.386 1.000, 0.421 − − School age (6–18) 21 7 33.3 14 66.7 2.29 0.022 0.618, 0.048 * − − − Adult (19–64) 30 14 46.7 16 53.3 0.52 0.605 0.319, 0.186 − − Senior (65–84) 8 4 50.0 4 50.0 0.00 1.000 0.490, 0.490 − Verbal trait history Speech 15 11 73.3 4 26.7 2.89 0.004 0.150, 0.783 * Language 17 8 47.1 9 52.9 0.34 0.731 0.394, 0.277 − − Reading 24 13 54.2 11 45.8 0.58 0.562 0.199, 0.365 − Spelling 18 10 55.6 8 44.4 0.67 0.502 0.214, 0.436 − Writing 3 1 33.3 2 66.7 0.87 0.386 1.000, 0.421 − − Participants scoring more than one 37 19 73.1 18 50.0 1.92 0.055 0.005, 0.469 SD below the mean in one or − more verbal trait domains a The row-wise percentages use the Total n in the second column as the denominator. The denominators for each percentage in the last row are 26 and 36, respectively b Minitab 17 Statistical Software (2010). [Computer software]. State College, PA: Minitab, Inc. (www.minitab.com) c * p < 0.05

Participants Table 2 Descriptives for VT and VT individuals in the family + − Tables 1, 2 describe demographic and phenotype vari- across the syllable repetition and nonword repetition tasks ables for participants meeting NWR task criteria for Verbal trait (VT ) Verbal trait (VT ) affected (Verbal Trait ; VT ) and not affected (Ver- + + − − + + SRT bal Trait ; VT ), including tests for significant differ- − − Mean (SD) 3.01 (3) 0.27 (0.7) ences between the proportions of each classification. − Skewness 0.96 0.25 The difference in the percentages of VT (41.9 %) com- − + pared to VT (58.1 %) participants was non-significant Kurtosis 0.05 0.53 − − (Z 1.82). Significantly fewer females met criteria for NRT = − VT (31.4 %) than VT (68.6 %; Z 3.35), but the Mean (SD) 2.07 (1.67) 0.42 (0.86) + − = − − proportion of males who met criteria for VT (55.6 %) Skewness 0.22 0.04 + − compared to the proportion who met criteria for VT Kurtosis 0.16 0.77 − − − (44.4 %) was non-significant (Z 0.82). Among the four = age groups, the only age group within which affection status differed significantly was the school-aged partici- domains lower than one standard deviation below stand- pants, who had a significantly lower percentage of par- ardized means, or any self- or parent-reported difficulty in ticipants who met criteria for VT (33.3 %) than VT any of the five domains (Supplemental Table 1). Of the Ver- + − (66.7 %; Z 2.29). bal Trait History variables in Table 1, only one verbal trait = − Last, Verbal Trait History for problems in verbal trait domain was associated with significant between group pro- domains of speech, language, reading, spelling, and/or writ- portions. A significantly greater percentage of participants ing were determined by test scores in any of the relevant with test scores, self-reported, or parental-reported histories

1 3 Hum Genet (2016) 135:1329–1341 1333 of speech disorders met the nonword criterion for VT (2) MAF 0.2–0.5 in the EUR reference population; (3) + (73.3 %) compared to the percentage who met criterion non-monomorphic marker within the pedigree;( 4) mini- for VT (26.7 %; Z 2.89). Using conventional criteria mum intermarker distance of 0.5 cM; and (5) restricted to − = for statistical significance, the percentage of VT partici- the 22 autosomes. A separate marker sub-panel was gener- + pants who had at least one test score or questionnaire entry ated for pedigree structure validation using similar criteria indicating a concern with any one of the five verbal traits as above except maximum LD threshold was equal to 0.25 (73.1 %) was not significantly larger than the percentage and MAF from 0.3 to 0.5 in the EUR reference population. of VT participants with such histories (50.0 %; Z 1.92; For genome-wide linkage analysis (excluding sex chromo- − = p 0.055; CI 0.005, 0.469). somes), a total of 5448, 5493, and 5498 markers for sub- = − panels 1, 2, and 3, respectively, were created. A sub-panel DNA collection and genotyping of 5454 genome-wide markers was created for pedigree structure validation. DNA was extracted from whole blood using the Gentra QC for appropriate parent–offspring relationships within Puregene Blood Kit (Qiagen) at the University of Nebraska the larger pedigree was assessed by comparing expected Medical Center. Genotyping across 551,839 single nucle- kinship coefficients (based on pedigree structure) and esti- otide polymorphism (SNP) markers was performed using mated coefficients computed by maximizing the likeli- the Illumina Infinium HumanCoreExome-24-v.1 at the Yale hood from available genotype data across the 5454 marker Center for Genome Analysis (Orange, CT). Genotypes sub-panel (Choi et al. 2009). Individual relationship pairs were called using Illumina GenomeStudio with a total of were flagged if the estimated kinship coefficient fell out- 547,644 (99.24 %) passing quality control (QC). One indi- side a 99.5 % confidence interval from expected. No sam- vidual failed QC due to low genotyping call rate and was ple swaps or incorrect parent–offspring relationships were excluded from the analysis. observed within the larger pedigree. From each marker sub-panel created for linkage analy- PBAP: marker sub‑selection, pedigree structure sis, PBAP prepared data files to generate IVs that described validation, and IBD computation for linkage analysis the flow of genetic information through a pedigree for an individual using gl_auto in the MORGAN suite of pro- Reference map files for the HumanCoreExome dense grams (Thompson 2011). The gl_auto program uses a marker panel were obtained from the Rutgers maps (Mat- combination of exact and Markov Chain Monte Carlo ise et al. 2007), with integrated linkage-physical maps in (MCMC) based estimations to sample IVs for each indi- sex averaged Haldane genetic distances (cM). Reference vidual. For the current analysis, IVs were sampled for each genotype data for Europeans were extracted from the main marker subpanel using the following parameters: 15,000 European (EUR) population data from the 1000 genomes MCMC burn-in iterations, sampling by scan and 100,000 project (The 1000 Genomes Project Consortium 2010) to MCMC iterations with progress checked every 20,000 determine linkage disequilibrium (LD) and minor allele fre- iterations (L-Sampler 0.2), saving 2000 realizations for = quencies (MAF) between markers for marker sub-selection. IV sampling. Sampled IVs were then converted to Sequen- We used the pedigree based analysis pipeline (PBAP) tial Oligonucleotide Linkage Analysis Routines (SOLAR) to sub-select genetic markers for pedigree quality con- (Almasy and Blangero 1998) compatible multipoint iden- trol (QC) and for interfacing with MORGAN (Thompson tity-by-descent (MIBD) matrices using custom scripts 2011) to calculate inheritance vectors (IV) used for link- written by the Wijsman lab, and imported into SOLAR for age analysis (Nato et al. 2015). Use of MORGAN allowed downstream linkage analysis. multipoint analysis on the complete pedigree. Generation of genome-wide SNP marker sub-panels from the dense Statistical analysis marker panel was conducted to (1) reduce LD between markers and minimize type 1 error, (2) reduce computa- All statistical analyses were conducted using the SOLAR tional time (while maximizing genotypic informativeness software package (version 7.3.9; Almasy and Blangero within the pedigree), and (3) perform QC on pedigree 1998). SOLAR utilizes a maximum likelihood vari- structure (i.e., parent–offspring swaps). PBAP marker sub- ance decomposition approach to estimate the influence of selection and pedigree structure validation are described in genetic and environmental effects on a phenotypic trait by detail elsewhere (Nato et al. 2015). Briefly, three non-over- modeling the covariance among family members relative to lapping marker sub-panels from the original dense marker genetic kinship (identity by descent). A liability threshold panel (Illumina HumanCoreExome-24-v.1) were gener- model was used to handle discrete traits under the assump- ated based on the following criteria: (1) maximum LD (r2) tion that the affection status of an individual was deter- threshold equal to 0.04 in the EUR reference population; mined by their underlying genetic risk exceeding a certain

1 3 1334 Hum Genet (2016) 135:1329–1341 threshold for the phenotype (i.e., VT ; Duggirala et al. linkage (SNP) markers on chromosome 13, encoding a total + 1997). Using maximum likelihood techniques, initial mod- of 41 genes (build GRCh37/hg19). To determine whether els were screened for the covariate effects of age, age2, sex, the linkage signal was due to an effect of pseudorandom age sex, age2 sex, and IQ. After covariate screening marker sub sampling, multipoint analyses on chromosome × × non-significant covariates (p > 0.1) were removed from the 13 were conducted again using non-overlapping marker final model. In addition, a variance component for house- subpanels 2 and 3, which were generated at the same time hold random effects that further controlled for shared envi- as marker subpanel 1 in PBAP. Findings were recapitulated ronment among nuclear families within the larger pedigree with peak LOD scores of 4.24 and 3.96 using marker sub- was included. The final model representing the log likeli- panels 2 and 3, respectively, again spanning the same region hood when the additive genetic variance was equal to 0 (no of chromosome 13q14.2–14.3 (Supplemental Table 3). Of linkage elements) and covarying for household and age*sex the 41 genes within base pair positions 48–53.5 Mb, 26 effects, was used as the null model for hypothesis testing genes show moderate expression in the developing brain during linkage analysis. (BrainSpan 2011), but only 8 have known neurological or Genome-wide multipoint variance component link- cognitive function (Carrozzo et al. 2007; de Bie et al. 2007; age analyses were conducted to examine linkage between Dening et al. 1989; Elpeleg et al. 2005; Hilschmann et al. VT and MIBDs. Multipoint linkage analysis considers 2002; Jaberi et al. 2013; Kind et al. 2014; La Piana et al. + recombination along a chromosome to determine the prob- 2016; Maas et al. 2015; Morris et al. 2012; Ocklenburg et al. ability that a trait locus is located within a genomic region. 2015; Ostergaard et al. 2007; Rice et al. 2007; Spiechow- Maximum likelihood estimates for linkage were calculated icz et al. 2006; Vidal et al. 1999, 2000; Wei and Hemmings at approximately 0.5 cM intervals across the 22 autosomes 2005; Xu et al. 2010; Yamagata et al. 1999; Yasuda et al. and compared against the null model (no linkage) using a 2007; Zhang et al. 2006; Supplemental Table 4). When likelihood ratio test (df 1). considering LOD scores greater than 3 (empirical p value = Empirical p values were computed using a simulation <0.0002), the linkage signal expands to 46–61 cM across 22 that generated a distribution of LOD scores under a null linkage markers spanning a 23.2 Mb region on chromosome model of no linkage. 1,000,000 simulations were con- 13q14.11-21.32, encompassing approximately 77 genes. ducted, each generating a random informative marker that Estimation of haplotypes in the extended family pro- was tested for linkage with VT status. This distribution vides evidence of at least 3 distinct haplotypes on chro- + of observed LOD scores in the simulation was then used mosome 13 segregating with VT (Fig. 2). A recombina- + to determine the empirical p value of the experimentally tion in Haplotype 1 (Haplotype 1-recombined) in affected observed LOD scores. individual 3 between SNPs rs7337528 and rs2981 defines Haplotypes were assigned using MERLIN (Abeca- the centromeric boundary of LOD 4.35 at 52 cM = sis et al. 2002). The large pedigree exceeded the bit limit (Figs. 2, 3). This particular segment of Haplotype 1, MERLIN could handle, thus the family was split into six defined by alleles shared IBD at rs2981, rs2812219, smaller subpedigrees for haplotype assignment and then rs6561602, and rs997687 (52.92–55.06 cM), segregates manually reconfigured to confirm consistency of haplo- with the phenotype in the lineage originated by a founder types called across the lineages. Five of the six subpedi- in the oldest generation. For the remaining information, grees were assigned based on distinct sublineages that data are presented without a conventional haplotype pedi- originated from Generation II, and included all individuals gree graphic to preserve anonymity of the family. Haplo- within the last four generations (Generations III–VI) of the type 2 originates from a married-in founder in Generation family. The sixth subpedigree consisted of all individuals in II and segregates to two siblings in Generation III, four the first two generations (Generations I and II) and select descendants in Generation IV and four descendants in individuals in Generations III and IV to confirm the trans- Generation V. Seven of ten descendants with Haplotype mission of the haplotypes observed in the aforementioned 2 are VT . Last, Haplotype 3 originates from a married- + five subpedigrees. in founder in Generation III, segregates to two siblings in generation IV, and four descendants in Generation V. Six of seven pedigree members with Haplotype 3 are VT . + Results There are no recombinations within Haplotypes 2 or 3 that would define the broader shoulders of the linkage Genome-wide multipoint linkage analyses for VT revealed signal from 45 to 62 cM (Fig. 3). Overall, the presence + 6 a peak LOD score of 4.35 (empirical p value <1 10− ) of at least three distinct haplotypes that cosegregate with × between 52 and 55 cM on chromosome 13q14.2–q14.3 VT —including married-in Haplotypes 2 and 3—impli- + (Fig. 1 and Supplemental Table 2) with marker subpanel 1. cate multiple contributing variants segregating through This region spans base pair positions 48–53.5 Mb across 5 this family.

1 3 Hum Genet (2016) 135:1329–1341 1335

Fig. 1 Multipoint linkage results conditioned on impaired NWR at chromosome 13. Genes and associated SNPs under the highest linkage peak of LOD 4.35, between 52 and 55 cM spanning physical positions 48–53.5 Mb on reference genome assembly build GRCh37/hg19 =

Fig. 2 Haplotypes spanning genomic location 45–62 cM on chromosome 13 segregating with Verbal Trait (affected) status in the family.+ The centromeric and telomeric boundaries of Haplotype 1-Recombined are defined by a recombinatorial events within individual 3 (Fig. 3). No recombinatorial events in the family offer clear centromeric and telomeric boundaries for Haplotypes 2 and 3

1 3 1336 Hum Genet (2016) 135:1329–1341

Fig. 3 Haplotype assignments spanning genomic location 45–62 cM the centromeric and telomeric border of the linkage signal spanning on chromosome 13. The pedigree depicted is truncated to reflect the 52–55 cM. Affected individuals are black diamonds, while unaffected recombinatorial event observed in affected individual 3 that outlines individuals are gray diamonds

Other suggestive linkage signals (LOD >1.5; empirical p Discussion value <0.01) were observed on chromosome 2q37.1, 4q12– 13.2, 4q25, 7q22.3–31.2, 8q24.3, and 12p13.33 (Supple- The present study identified a linkage signal spanning mental Table 2). Most notably, the linkage peak spanning chromosome 13q14–q21 using a categorical phenotype 7q22.3–q31.2 has a max LOD 2.06 and contains the (VT or VT ) derived from performance on NWR in an = + − gene FOXP2—a causal gene for Childhood Apraxia of extended pedigree with a history of verbal trait disorders. Speech (Fisher et al. 1998; Lai et al. 2001). This region encompasses SLI3 on chromosome 13q21,

1 3 Hum Genet (2016) 135:1329–1341 1337 a SLI locus previously identified by Bartlett et al. (2002), Danish Dementia with similar pathology to Alzheimer using a family-based linkage analysis in five Canadian disease (Vidal et al. 1999, 2000). Setdb2, the zebrafish families of Celtic ancestry with a history of specific lan- ortholog of SETDB2, is known to regulate left–right asym- guage impairment (SLI). Their analysis was conditioned on metry in the zebrafish central nervous system (Xu et al. a categorical reading-IQ discrepancy phenotype (nonword 2010). Human epidemiological research has also associated reading score at least one standard deviation below perfor- SETDB2 with handedness (left versus right preference) mance IQ), which they replicated in a larger independent with specific variants linked to reduction in laterality (Ock- US sample using the same phenotype (Bartlett et al. 2004). lenburg et al. 2015). This provides an interesting parallel This region has also been implicated in autism spectrum to previous evidence that suggests language and reading disorder (ASD), a neurodevelopmental disorder with a core disability are linked to atypical cerebral laterality (asym- language component in combination with other core abnor- metry) since language-related behavior is typically left malities in social and repetitive behaviors. A linkage signal lateralized (Leonard and Eckert 2008; Scerri et al. 2011). at chromosome 13q21 was observed with a language delay However, it is important to note that SETDB2 has not yet phenotype in the Collaborative Linkage Study of Autism been directly implicated in reading or language disability. (CLSA; Bradford et al. 2001). Furthermore, deletions at ATP7B is a known gene associated with Wilson disease, 13q12 through 13q21 have been reported in three subjects which is a disorder characterized by the deposition of cop- with ASD and poor receptive and expressive vocabulary per in the liver, brain, and other tissues, leading to neuro- (but normal speech), and in a subject with ASD with audi- logical and cognitive deterioration including memory loss, tory processing deficits (Mitter et al. 2011; Smith et al. tremors, and emotional changes (de Bie et al. 2007; Den- 2002; Steele et al. 2001). These deleted segments partially ing Tr 1989). PCDH8 is part of the protocadherin family overlap the linkage peak in our present study, where LOD of CNS-specific cell adhesion molecules that plays a role >1.5 (Supplemental Tables 2 and 3). The convergence of in the development of neural circuitry (Hilschmann et al. these findings associated with 13q14–q21 with related lan- 2002; Yamagata et al. 1999). Interestingly, the rat ortholog guage phenotypes that underlie verbal trait disorders pro- of PCDH8, Arcadlin, has been implicated in synaptic func- vides compelling support for this locus. tion and is dynamically expressed upon activation of hip- Within this pedigree, there are at least three distinct hap- pocampal circuitry—a neural network necessary for learn- lotypes segregating with VT , of which, only Haplotype ing and memory (Yasuda et al. 2007). + 1 originated with a founder in the oldest generation—the By examining the extended linkage peak spanning other two are more recently married into, consistent with 45–62 cM with LOD scores >3, we identified another assortative mating. Within the EUR reference population of three genes in the protocadherin family located telemetric 1000 genomes project, Haplotype 1-Recombined (Fig. 2) is to PCDH8—PCDH17, PCDH20, and PCDH—each of common with a frequency of 0.046. In a clinical context, which encodes cell–cell adhesion molecules that are pri- verbal trait disorders such as developmental dyslexia and marily expressed in the brain (Kim et al. 2010). Variation in specific language impairment (SLI) have a high prevalence PCDH9 has been linked to ASD and SLI, while PCDH17 in the United States. The prevalence of developmental dys- is highly expressed in the prefrontal and anterior regions lexia is 7 % in the general population, and the prevalence of of the temporal cortex and subcortical structures such as specific language impairment (SLI) is 5–8 % among pre- the thalamus, ventral striatum, and anterior cingulate—an school children (Peterson and Pennington 2015; Tomblin expression pattern highlighting an overlap with corticos- et al. 1997). Different haplotypes segregating within the triatothalamic circuitry critical for higher order cognitive family could indicate a single gene with different causal function and language development (Abrahams et al. 2007; variants segregating within the family. It is also possible Marshall et al. 2008). that different genes at the same locus, or at different loci, Our linkage findings on chromosome 13q do not corre- are mediating NWR performance. Further fine mapping spond to other genome-wide linkage scans conditioned on and sequencing of the region is necessary to disentangle NWR. A family-based linkage analysis conducted by the these possibilities and elucidate the potential variants driv- SLI consortium localized to chromosome 16q with further ing the signal observed in the present study. fine mapping identifying CMIP and ATP2C2 as potential Underneath the peak linkage signal spanning the gene candidates mediating NWR in their sample (SLI Con- 52–55 cM region of chromosome 13 with LOD >4, there sortium 2002, 2004; Newbury et al. 2009). Another study are interesting gene candidates with known function in performed by Brkanac et al. (2008), observed linkage sig- neuropsychiatric disorders and neurodevelopment. ITM2B nals on chromosomes 4p12, 12p, and 17q in families with encodes a transmembrane protein that helps to inhibit the a history of dyslexia. Discrepancies in genomic regions accumulation of beta-amyloid, but mutations have been associated with NWR, in part, may be due to differences implicated in Familial British Dementia and Familial in the particular NWR test used to evaluate respective

1 3 1338 Hum Genet (2016) 135:1329–1341 subjects. Although the measures used in these studies do transcription reliability (Shriberg et al. 2009). Because ostensibly assess NWR, there are differences in each that individuals tested in the extended pedigree ranged in age may more heavily tap into different combinations of under- from 3 to 95 years old, it would not be optimal to test all lying cognitive and/or behavioral abilities, such as phono- individuals on only the NRT or SRT, as they may differ logical working memory, long-term lexical knowledge, and in their sensitivity to persistent types and levels of NWR articulatory difficulty with nonsense words (Estes et al. deficits across the lifespan. 2007; Gathercole 1995). Ascertainment differences and In conclusion, we found a statistically significant differences in age ranges between the present study and genome-wide multipoint linkage signal on chromosome others could also contribute to the observed discrepancies. 13q14–q21 using a NWR phenotype in an extended pedi- The SLI consortium used a family-based linkage analysis gree with a family history of verbal trait disorder. We examining 186 nuclear families affected with SLI (SLI hypothesize that the region of 13q14–q21 is a susceptibility Consortium 2002, 2004). Brkanac et al. (2008), used a fam- locus for verbal trait disorders, but additional work must be ily-based design examining 144 families with a history of conducted to (1) identify the gene(s) in this region contrib- dyslexia, whereas the present study examined one extended uting to the linkage signals observed in the present study family with verbal trait disorder that could be derived from and others that have been identified this same region, and one or more rare variants. The small number of subjects in (2) elucidate the complex genetic and environmental inter- each of these studies would significantly limit the power actions that may increase susceptibility. to detect rare and uncommon variants. Ultimately, these findings may also reflect locus heterogeneity and highlight Acknowledgments This study was supported by grant National Insti- different molecular and biological mechanisms associated tutes of Health (NIH) R01 DC000496 (LDS), NIH HD03352 (Wais- man Center), MH 094293 (EMW, AQN), NIH R01 N5043530 (JRG), with NWR. P50 HD027802 (JRG), T32 HD07094 (DTT), and F31 DC012270 Due to the complex inheritance pattern of impaired (JDE). We thank our colleagues in the Phonology Project, Wais- NWR performance, a nonparametric analysis using vari- man Center, for their assistance, and the staff at the Yale Center for ance components was used so that pre-specified values Genome Analysis for genotyping services We extend our sincere grat- itude to a family member whose assistance and insights were central for parameters defining the genetic model would not be to this research, and to all participants for their contribution to this required (Bailey-Wilson 2004). This is in contrast to a study. parametric analysis that requires the specification of a genetic model, with the concern that a poorly specified Compliance with ethical standards model could lead to suboptimal results. An advantage to Conflict of interest The authors declare that they have no conflict of using a variance components approach is that it tends to interest. be more powerful relative to other trait mapping methods (Kleensang et al. 2010). However, a limitation is that Open Access This article is distributed under the terms of the Crea- variance components provides poorer localization of the tive Commons Attribution 4.0 International License (http://crea- tivecommons.org/licenses/by/4.0/), which permits unrestricted use, trait locus compared to a parametric analysis, and gener- distribution, and reproduction in any medium, provided you give ally requires additional fine mapping to isolate the region appropriate credit to the original author(s) and the source, provide a (Amos and de Andrade 2001; Williams et al. 1997). An link to the Creative Commons license, and indicate if changes were additional limitation is that we used a composite vari- made. able across two different NWR tests—performing more than one standard deviation below the age–sex standardized mean on either the NRT or SRT—to derive VT References + or status. Although both measures evaluate NWR, the − individual test items differ. As described previously, the Abecasis GR, Cherny SS, Cookson WO, Cardon LR (2002) Merlin SRT focuses on the repetition of syllables that comprise rapid analysis of dense genetic maps using sparse gene flow trees. Nat Genet 30:97. doi:10.1038/ng786 only four “early-8” consonants and can be used to exam- Abrahams BS, Tentler D, Perederiy JV, Oldham MC, Coppola G, ine NWR ability in young children and speakers of any Geschwind DH (2007) Genome-wide analyses of human peri- age with limited phonetic inventories or speech sound sylvian cerebral cortical patterning. Proc Natl Acad Sci USA disorder (Shriberg et al. 2009). In comparison, although 104:17849–17854. doi:10.1073/pnas.0706128104 Alarcón M, Cantor RM, Liu J, Gilliam TC, Geschwind DH (2002) the NRT was designed to exclude late developing English Evidence for a language quantitative trait locus on chromosome consonants (the ‘late-8’), younger children and speakers 7q in multiplex autism families. Am J Hum Genet 70:60–71. of any age with speech sound disorder can have articu- doi:10.1086/338241 lation errors repeating the 9 different vowels and diph- Alarcón M, Yonan AL, Gilliam TC, Cantor RM, Geschwind DH (2005) Quantitative genome scan and ordered-subsets analysis of thongs, and 11 different consonants in the nonsense autism endophenotypes support language QTLs. Mol Psychiatry words, thus confounding test performance and reducing 10:747–757. doi:10.1038/sj.mp.4001666

1 3 Hum Genet (2016) 135:1329–1341 1339

Alarcón M, Abrahams BS, Stone JL, Duvall JA, Perederiy JV, Bomar Choi Y, Wijsman EM, Weir BS (2009) Case-control association test- JM, Sebat J, Wigler M, Martin CL, Ledbetter DH, Nelson SF, ing in the presence of unknown relationships. Genet Epidemiol Cantor RM, Geschwind DH (2008) Linkage, association, and 33:668–678. doi:10.1002/gepi.20418 gene-expression analyses identify CNTNAP2 as an autism-sus- de Bie P, Muller P, Wijmenga C, Klomp LWJ (2007) Molecular patho- ceptibility gene. Am J Hum Genet 82:150–159. doi:10.1016/j. genesis of Wilson and Menkes disease: correlation of mutations ajhg.2007.09.005 with molecular defects and disease phenotypes. J Med Genet Almasy L, Blangero J (1998) Multipoint quantitative-trait linkage 44:673–688. doi:10.1136/jmg.2007.052746 analysis in general pedigrees. Am J Hum Genet 62:1198–1211. Dening Tr BGE (1989) Wilson’s disease: psychiatric symptoms in doi:10.1086/301844 195 cases. Arch Gen Psychiatry 46:1126–1134. doi:10.1001/arc American Speech-Language-Hearing Association (2007) Ad Hoc hpsyc.1989.01810120068011 Committee on Childhood Apraxia of Speech: Childhood Apraxia Dollaghan C, Campbell TF (1998) Nonword repetition and child of Speech [Position Statement]. ASHA, Rockville language impairment. J Speech Lang Hear Res 41:1136–1146. Amos CI, de Andrade M (2001) Genetic linkage methods doi:10.1044/jslhr.4105.1136 for quantitative traits. Stat Methods Med Res 10:3–25. Duggirala R, Williams JT, Williams-Blangero S, Blangero J (1997) A doi:10.1177/096228020101000102 variance component approach to dichotomous trait linkage analysis Archibald LMD, Gathercole SE (2006) Nonword repetition: a using a threshold model. Genet Epidemiol 14:987–992. doi:10.1002/ comparison of tests. J Speech Lang Hear Res 49:970–983. (SICI)1098-2272(1997)14:6<987:AID-GEPI71>3.0.CO;2-G doi:10.1044/1092-4388(2006/070) Elpeleg O, Miller C, Hershkovitz E, Bitner-Glindzicz M, Bondi- Bailey-Wilson JE (2004) Parametric versus nonparametric and two- Rubinstein G, Rahman S, Pagnamenta A, Eshhar S, Saada point versus multipoint: controversies in gene mapping. Ency- A (2005) Deficiency of the ADP-forming succinyl-CoA syn- clopedia of Genetics, Genomics, Proteomics and Bioinformatics. thase activity is associated with encephalomyopathy and mito- Wiley, Hoboken chondrial DNA depletion. Am J Hum Genet 76:1081–1086. Bartlett CW, Flax JF, Logue MW, Vieland VJ, Bassett AS, Tallal P, doi:10.1086/430843 Brzustowicz LM (2002) A major susceptibility locus for spe- Estes KG, Evans JL, Else-Quest N (2007) Differences in the nonword cific language impairment is located on 13q21. Am J Hum Genet repetition performance of children with and without specific 71:45–55. doi:10.1086/341095 language impairment: a meta-analysis. J Speech Lang Hear Res Bartlett CW, Flax JF, Logue MW, Smith BJ, Vieland VJ, Tallal P, 50:177–195. doi:10.1044/1092-4388(2007/015) Brzustowicz LM (2004) Examination of potential overlap in Fisher SE, Vargha-Khadem F, Watkins KE, Monaco AP, Pembrey autism and language loci on chromosomes 2, 7, and 13 in two ME (1998) Localisation of a gene implicated in a severe speech independent samples ascertained for specific language impair- and language disorder. Nat Genet 18:168–170. doi:10.1038/ ment. Hum Hered 57:10–20. doi:10.1159/000077385 ng0298-168 Bishop DVM, North T, Donlan C (1996) Nonword repetition as a Gathercole SE (1995) Is nonword repetition a test of phonological behavioural marker for inherited language impairment: evidence memory or long-term knowledge–it all depends on the non- from a Twin Study. J Child Psychol Psychiatry 37:391–403. words. Mem Cognit 23:83–94. doi:10.3758/Bf03210559 doi:10.1111/j.1469-7610.1996.tb01420.x Glahn DC, Knowles EE, McKay DR, Sprooten E, Raventós H, Blan- Bishop DVM, Adams CV, Norbury CF (2004) Using nonword rep- gero J, Gottesman II, Almasy L (2014) Arguments for the sake of etition to distinguish genetic and environmental influences on endophenotypes: examining common misconceptions about the early literacy development: a study of 6-year-old twins. Am J use of endophenotypes in psychiatric genetics. Am J Med Genet Med Genet B Neuropsychiatr Genet 129B:94–96. doi:10.1002/ B Neuropsychiatr Genet. doi:10.1002/ajmg.b.32221 ajmg.b.30065 Goldman R, Fristoe M (2000) Goldman-Fristoe test of articulation, Bradford Y, Haines J, Hutcheson H, Gardiner M, Braun T, Sheffield 2nd edn. American Guidance Service Inc, Circle Pines V, Cassavant T, Huang W, Wang K, Vieland V, Folstein S, San- Hilschmann N, Barnikol HU, Barnikol-Watanabe S, Götz H, Kratzin tangelo S, Piven J (2001) Incorporating language phenotypes H, Thinnes FP (2002) The immunoglobulin-like genetic pre- strengthens evidence of linkage to autism. Am J Med Genet determination of the brain: the protocadherins, blueprint of the 105:539–547. doi:10.1002/ajmg.1497 neuronal network. Naturwissenschaften 88:2–12. doi:10.1007/ BrainSpan: Atlas of the Developing Human Brain [Internet]. Funded s001140000190 by ARRA Awards 1RC2MH089921-01, 1RC2MH090047-01, Hurst JA, Baraitser M, Auger E, Graham F, Norell S (1990) An and 1RC2MH089929-01. © 2011. http://developinghumanbrain. extended family with a dominantly inherited speech disorder. Dev org Med Child Neurol 32:352–355. doi:10.1111/j.1469-8749.1990. Brkanac Z, Chapman NH, Igo RP Jr, Matsushita MM, Nielsen K, tb16948.x Berninger VW, Wijsman EM, Raskind WH (2008) Genome scan Jaberi E, Chitsazian F, Ali Shahidi G, Rohani M, Sina F, Safari I, of a nonword repetition phenotype in families with dyslexia: evi- Malakouti Nejad M, Houshmand M, Klotzle B, Elahi E (2013) dence for multiple loci. Behav Genet 38:462–475. doi:10.1007/ The novel mutation p.Asp251Asn in the [beta]-subunit of succi- s10519-008-9215-2 nate-CoA ligase causes encephalomyopathy and elevated succi- Campbell TF, Dollaghan CA, Rockette HE, Paradise JL, Feldman nylcarnitine. J Hum Genet 58:526–530. doi:10.1038/jhg.2013.45 HM, Shriberg LD, Sabo DL, Kurs-Lasky M (2003) Risk factors Kang C, Drayna D (2011) Genetics of speech and language disor- for speech delay of unknown origin in 3-year-old children. Child ders. Annu Rev Genomics Hum Genet 12:145–164. doi:10.1146/ Dev 74:346–357. doi:10.1111/1467-8624.7402002 annurev-genom-090810-183119 Carrozzo R, Dionisi-Vici C, Steuerwald U, Lucioli S, Deodato F, Di Kaufman AS, Kaufman NL (2004) Kaufman brief intelligence test, Giandomenico S, Bertini E, Franke B, Kluijtmans LAJ, Mes- Second edn (KBIT-2), Pearson Assessments, San Antonio chini MC, Rizzo C, Piemonte F, Rodenburg R, Santer R, San- Kim S-Y, Yasuda S, Tanaka H, Yamagata K, Kim H (2010) Non- torelli FM, van Rooij A, Vermunt-de Koning D, Morava E, clustered protocadherin. Cell Adh Migr 5:97–105. doi:10.4161/ Wevers RA (2007) SUCLA2 mutations are associated with mild cam.5.2.14374 methylmalonic aciduria, Leigh-like encephalomyopathy, dysto- Kind B, Muster B, Staroske W, Herce HD, Sachse R, Rapp A, nia and deafness. Brain 130:862–874 Schmidt F, Koss S, Cardoso MC, Lee-Kirsch MA (2014) Altered

1 3 1340 Hum Genet (2016) 135:1329–1341

spatio-temporal dynamics of RNase H2 complex assembly at Newbury DF, Winchester L, Addis L, Paracchini S, Buckingham L-L, replication and repair sites in Aicardi-Goutières syndrome. Hum Clark A, Cohen W, Cowie H, Dworzynski K, Everitt A, Goodyer Mol Genet 23:5950–5960 IM, Hennessy E, Kindley AD, Miller LL, Nasir J, O’Hare A, Kleensang A, Franke D, Alcaïs A, Abel L, Müller-Myhsok B, Ziegler Shaw D, Simkin Z, Simonoff E, Slonims V, Watson J, Ragoussis A (2010) An extensive comparison of quantitative trait loci map- J, Fisher SE, Seckl JR, Helms PJ, Bolton PF, Pickles A, Conti- ping methods. Hum Hered 69:202–211. doi:10.1159/000289596 Ramsden G, Baird G, Bishop DVM, Monaco AP (2009) CMIP La Piana R, Uggetti C, Roncarolo F, Vanderver A, Olivieri I, Tonduti and ATP2C2 modulate phonological short-term memory in lan- D, Helman G, Balottin U, Fazzi E, Crow YJ, Livingston J, Orcesi guage impairment. Am J Hum Genet 85:264–272. doi:10.1016/j. S (2016) Neuroradiologic patterns and novel imaging findings in ajhg.2009.07.004 Aicardi-Goutières syndrome. Neurology 86:28–35. doi:10.1212/ Ocklenburg S, Arning L, Gerding WM, Hengstler JG, Epplen JT, WNL.0000000000002228 Güntürkün O, Beste C, Akkad DA (2015) Left-right axis differ- Lai CSL, Fisher SE, Hurst JA, Vargha-Khadem F, Monaco AP (2001) entiation and functional lateralization: a haplotype in the methyl- A forkhead-domain gene is mutated in a severe speech and lan- transferase encoding gene SETDB2 might mediate handedness in guage disorder. Nature 413:519–523. doi:10.1038/35097076 healthy adults. Mol Neurobiol. doi:10.1007/s12035-015-9534-2 Lenzenweger MF (2013) Endophenotype, intermediate phenotype, Ostergaard E, Hansen FJ, Sorensen N, Duno M, Vissing J, Larsen PL, biomarker: definitions, concept comparisons, clarifications. Faeroe O, Thorgrimsson S, Wibrand F, Christensen E, Schwartz Depress Anxiety 30(3):185–189 M (2007) Mitochondrial encephalomyopathy with elevated Leonard CM, Eckert MA (2008) Asymmetry and dyslexia. Dev Neu- methylmalonic acid is caused by SUCLA2 mutations. Brain ropsychol 33:663–681. doi:10.1080/87565640802418597 130:853–861 Lewis BA, Freebairn LA, Taylor HG (2000) Academic outcomes in Peterson RL, Pennington BF (2015) Developmental dys- children with histories of speech sound disorders. J Commun lexia. Annu Rev Clin Psychol 11:283–307. doi:10.1146/ Disord 33:11–30. doi:10.1016/S0021-9924(99)00023-4 annurev-clinpsy-032814-112842 Lohmeier HL, Shriberg LD (2011) Reference data for the syllable Potter NL, Hall S, Karlsson HB, Fourakis M, Lohmeier HL, repetition task (SRT). Technical Report No. 17, Phonology Pro- McSweeny JL, Tilkens CM, Wilson DL, Shriberg LD (2012) ject, Waisman Center, University of Wisconsin-Madison Reference data for the Madison Speech Assessment Protocol Maas RR, Marina AD, Brouwer APM, Wevers RA, Rodenburg RJ, (MSAP): a database of 150 participants 3-to-18 years of age with Wortmann SB (2015) SUCLA2 Deficiency: a deafness-dystonia typical speech. Technical Report No. 18, Phonology Project, syndrome with distinctive metabolic findings (Report of a New Waisman Center, University of Wisconsin-Madison Patient and Review of the Literature). Springer Berlin Heidel- Rice G, Patrick T, Parmar R, Taylor CF, Aeby A, Aicardi J, Artuch berg, p 1–6. doi: 10.1007/8904_2015_464 R, Montalto SA, Bacino CA, Barroso B, Baxter P, Benko WS, MacDermot KD, Bonora E, Sykes N, Coupe A-M, Lai CSL, Vernes Bergmann C, Bertini E, Biancheri R, Blair EM, Blau N, Bon- SC, Vargha-Khadem F, McKenzie F, Smith RL, Monaco AP, thron DT, Briggs T, Brueton LA, Brunner HG, Burke CJ, Carr Fisher SE (2005) Identification of FOXP2 truncation as a novel IM, Carvalho DR, Chandler KE, Christen H-J, Corry PC, Cowan cause of developmental speech and language deficits. Am J Hum FM, Cox H, D’Arrigo S, Dean J, De Laet C, De Praeter C, Déry Genet 76:1074–1080. doi:10.1086/430841 C, Ferrie CD, Flintoff K, Frints SGM, Garcia-Cazorla A, Gener Marshall CR, Noor A, Vincent JB, Lionel AC, Feuk L, Skaug J, Shago B, Goizet C, Goutières F, Green AJ, Guët A, Hamel BCJ, Hay- M, Moessner R, Pinto D, Ren Y, Thiruvahindrapduram B, Fie- ward BE, Heiberg A, Hennekam RC, Husson M, Jackson AP, big A, Schreiber S, Friedman J, Ketelaars CE, Vos YJ, Ficicioglu Jayatunga R, Jiang Y-H, Kant SG, Kao A, King MD, Kingston C, Kirkpatrick S, Nicolson R, Sloman L, Summers A, Gibbons HM, Klepper J, van der Knaap MS, Kornberg AJ, Kotzot D, CA, Teebi A, Chitayat D, Weksberg R, Thompson A, Vardy C, Kratzer W, Lacombe D, Lagae L, Landrieu PG, Lanzi G, Leitch Crosbie V, Luscombe S, Baatjes R, Zwaigenbaum L, Roberts W, A, Lim MJ, Livingston JH, Lourenco CM, Lyall EGH, Lynch Fernandez B, Szatmari P, Scherer SW (2008) Structural variation SA, Lyons MJ, Marom D, McClure JP, McWilliam R, Melancon of chromosomes in autism spectrum disorder. Am J Hum Genet SB, Mewasingh LD, Moutard M-L, Nischal KK, Østergaard JR, 82:477–488. doi:10.1016/j.ajhg.2007.12.009 Prendiville J, Rasmussen M, Rogers RC, Roland D, Rosser EM, Matise TC, Chen F, Chen W, De La Vega FM, Hansen M, He C, Rostasy K, Roubertie A, Sanchis A, Schiffmann R, Scholl-Bürgi Hyland FCL, Kennedy GC, Kong X, Murray SS, Ziegle JS, S, Seal S, Shalev SA, Corcoles CS, Sinha GP, Soler D, Spiegel Stewart WCL, Buyske S (2007) A second-generation combined R, Stephenson JBP, Tacke U, Tan TY, Till M, Tolmie JL et al linkage–physical map of the human genome. Genome Res (2007) Clinical and molecular phenotype of Aicardi-Goutières 17:1783–1786. doi:10.1101/gr.7156307 syndrome. Am J Hum Genet 81:713–725. doi:10.1086/521373 Mitter D, Ullmann R, Muradyan A, Klein-Hitpas L, Kanber D, Ounap Rice GM, Raca G, Jakielski KJ, Laffin JJ, Iyama-Kurtycz CM, Hart- K, Kaulisch M, Lohmann D (2011) Genotype-phenotype corre- ley SL, Sprague RE, Heintzelman AT, Shriberg LD (2012) Phe- lations in patients with retinoblastoma and interstitial 13q dele- notype of FOXP2 haploinsufficiency in a mother and son. Am J tions. Eur J Hum Genet 19:947–958. doi:10.1038/ejhg.2011.58 Med Genet A 158A:174–181. doi:10.1002/ajmg.a.34354 Moore MW, Tompkins CA, Dollaghan CA (2010) Manipulating artic- Royal College of Speech and Language Therapists (2011) Develop- ulatory demands in non-word repetition: a ‘late-8’ non-word rep- mental verbal dyspraxia policy statement. The Royal College of etition task. Clin Linguist Phon 24:997–1008. doi:10.3109/0269 Speech and Language Therapists, London 9206.2010.510917 Scerri TS, Brandler WM, Paracchini S, Morris AP, Ring SM, Richard- Morris CP, Baune BT, Domschke K, Arolt V, Swagell CD, Hughes son AJ, Talcott JB, Stein J, Monaco AP (2011) PCSK6 is asso- IP, Lawford BR, McD Young R, Voisey J (2012) KPNA3 varia- ciated with handedness in individuals with dyslexia. Hum Mol tion is associated with schizophrenia, major depression, opiate Genet 20:608–614. doi:10.1093/hmg/ddq475 dependence and alcohol dependence. Dis Markers. doi:10.3233/ Scheer-Cohen AR, Holt AS, Karlsson HB, Mabie HL, McSweeny JL, dma-2012-0921 Tilkens CM, Shriberg LD (2013) Reference data for the Madi- Nato AQ, Chapman NH, Sohi HK, Nguyen HD, Brkanac Z, Wijsman son Speech Assessment Protocol (MSAP): a database of fifty EM (2015) PBAP: a pipeline for file processing and quality con- 20-to-80 year old participants with typical speech, Technical trol of pedigree data with dense genetic markers. Bioinformatics Report No. 20, Phonology Project, Waisman Center, University 31:3790–3798. doi:10.1093/bioinformatics/btv444 of Wisconsin-Madison

1 3 Hum Genet (2016) 135:1329–1341 1341

Semel E, Wiig EH, Secord WA (2004) Clinical evaluation of language Vernes SC, Newbury DF, Abrahams BS, Winchester L, Nicod J, fundamentals, Fourth Edition—Screening test (CELF-4 screen- Groszer M, Alarcón M, Oliver PL, Davies KE, Geschwind DH, ing test). The Psychological Corporation, Toronto Monaco AP, Fisher SE (2008) A functional genetic link between Shriberg LD (1993) Four new speech and prosody-voice measures distinct developmental language disorders. N Engl J Med for genetics research and other studies in developmental pho- 359:2337–2345. doi:10.1056/NEJMoa0802828 nological disorders. J Speech Hear Res 36:105. doi:10.1044/ Vidal R, Frangione B, Rostango A, Mead S, Revesz T, Plant G, Ghiso jshr.3601.105 J (1999) A stop-codon mutation in the BRI gene associated with Shriberg LD, Lohmeier HL (2008) The Syllable Repetition Task familial British dementia. Nature 399:776. doi:10.1038/21637 (SRT). Technical Report No. 14, Phonology Project, Waisman Vidal R, Révész T, Rostagno A, Kim E, Holton JL, Bek T, Bojsen- Center, University of Wisconsin-Madison Møller M, Braendgaard H, Plant G, Ghiso J, Frangione B (2000) Shriberg LD, Tomblin JB, McSweeny JL (1999) Prevalence of A decamer duplication in the 3’ region of the BRI gene originates speech delay in 6-year-old children and comorbidity with lan- an amyloid peptide that is associated with dementia in a Danish guage impairment. J Speech Lang Hear Res 42:1461–1481. kindred. Proc Natl Acad Sci USA 97:4920–4925. doi:10.1073/ doi:10.1044/jslhr.4206.1461 pnas.080076097 Shriberg LD, Ballard KJ, Duffy JR, Odell KH, Williams CA (2006) Wei J, Hemmings GP (2005) The KPNA3 gene may be a suscepti- Speech, prosody, and voice characteristics of a mother and bility candidate for schizophrenia. Neurosci Res 52:342–346. daughter with a 7;13 translocation affecting FOXP2. J Speech doi:10.1016/j.neures.2005.04.005 Lang Hear Res 49:500–525. doi:10.1044/1092-4388(2006/038) Wiig EH, Secord WA, Semel E (2004) Clinical evaluation of language Shriberg LD, Lohmeier HL, Campbell TF, Dollaghan CA, fundamentals—Preschool, Second edn (CELF Preschool-2). The Green JR, Moore CA (2009) A nonword repetition task Psychological Corporation, Toronto for speakers with misarticulations: the syllable repeti- Wijsman EM, Peterson D, Leutenegger A-L, Thomson JB, Goddard tion task (SRT). J Speech Lang Hear Res 52:1189–1212. KAB, Hsu L, Berninger VW, Raskind WH (2000) Segregation doi:10.1044/1092-4388(2009/08-0047) analysis of phenotypic components of learning disabilities. I. Shriberg LD, Lohmeier HL, Strand EA, Jakielski KJ (2012) Encod- nonword memory and digit span. Am J Hum Genet 67:631–646. ing, memory, and transcoding deficits in childhood apraxia of doi:10.1086/303044 speech. Clin Linguist Phon 26:445–482. doi:10.3109/02699206. Williams JT, Duggirala R, Blangero J (1997) Statistical prop- 2012.655841 erties of a variance components method for quantita- SLI Consortium (2002) A genome-wide scan identifies two novel loci tive trait linkage analysis in nuclear families and extended involved in specific language impairment*. Am J Hum Genet pedigrees. Genet Epidemiol 14:1065–1070. doi:10.1002/ 70:384–398. doi:10.1086/338649 (SICI)1098-2272(1997)14:6<1065:AID-GEPI84>3.0.CO;2-F SLI Consortium (2004) Highly significant linkage to the SLI1 Woodcock RW, McGrew K, Mather N (2001) Woodcock-Johnson locus in an expanded sample of individuals affected by spe- tests of achievement, 3rd edition (WJ-III). Riverside Publishing, cific language impairment. Am J Hum Genet 74:1225–1238. Rolling Meadows doi:10.1086/421529 Wren YE, Roulstone SE, Miller LL (2012) Distinguishing groups of Smith M, Woodroffe A, Smith R, Holguin S, Martinez J, Filipek PA, children with persistent speech disorder: findings from a pro- Modahl C, Moore B, Bocian ME, Mays L, Laulhere T, Flodman spective population study. Logoped Phoniatr Vocol 37:1–10. doi: P, Spence MA (2002) Molecular genetic delineation of a deletion 10.3109/14015439.2011.625973 of chromosome 13q12 q13 in a patient with autism and audi- Xu P-F, Zhu K-Y, Jin Y, Chen Y, Sun X-J, Deng M, Chen S-J, Chen tory processing deficits.→ Cytogenet Genome Res 98:233–239. Z, Liu TX (2010) Setdb2 restricts dorsal organizer territory and doi:10.1159/000071040 regulates left–right asymmetry through suppressing fgf8 activ- Spiechowicz M, Bernstein H-G, Dobrowolny H, Les´niak W, Maw- ity. Proc Natl Acad Sci USA 107:2521–2526. doi:10.1073/ rin C, Bogerts B, Kuz´nicki J, Filipek A (2006) Density of pnas.0914396107 Sgt1-immunopositive neurons is decreased in the cerebral cor- Yamagata K, Andreasson KI, Sugiura H, Maru E, Dominique M, Irie tex of Alzheimer’s disease brain. Neurochem Int 49:487–493. Y, Miki N, Hayashi Y, Yoshioka M, Kaneko K, Kato H, Wor- doi:10.1016/j.neuint.2006.03.009 ley PF (1999) Arcadlin is a neural activity-regulated cadherin Steele MM, Al-Adeimi M, Siu VM, Fan Y-S (2001) Brief report: a involved in long term potentiation. J Biol Chem 274:19473– case of autism with interstitial deletion of chromosome 13. J 19479. doi:10.1074/jbc.274.27.19473 Autism Dev Disord 31:231–234. doi:10.1023/A:1010759401344 Yasuda S, Tanaka H, Sugiura H, Okamura K, Sakaguchi T, Tran U, The 1000 Genomes Project Consortium (2010) A map of human Takemiya T, Mizoguchi A, Yagita Y, Sakurai T, De Robertis EM, genome variation from population-scale sequencing. Nature Yamagata K (2007) Activity-induced protocadherin arcadlin 467:1061–1073. doi:10.1038/nature09534 regulates dendritic spine number by triggering N-cadherin endo- Thompson E (2011) The structure of genetic linkage data: from LIPED cytosis via TAO2β and p38 MAP kinases. Neuron 56:456–471. to 1 M SNPs. Hum Hered 71:86–96. doi:10.1159/000313555 doi:10.1016/j.neuron.2007.08.020 Tomblin JB, Records NL, Buckwalter P, Zhang X, Smith E, O’Brien Zhang H, Ju G, Wei J, Hu Y, Liu L, Xu Q, Chen Y, Sun Z, Liu S, Yu M (1997) Prevalence of specific language impairment in kinder- Y, Guo Y, Shen Y (2006) A combined effect of the KPNA3 and garten children. J Speech Lang Hear Res 40:1245. doi:10.1044/ KPNB3 genes on susceptibility to schizophrenia. Neurosci Lett jslhr.4006.1245 402:173–175. doi:10.1016/j.neulet.2006.03.073

1 3